Hallucination Evaluation on CRPE relation
79.1AccuracyInternVL2.5-26B
Evaluation Results
| Method | Links | |
|---|---|---|
| InternVL2.5-26BModel Name=InternVL2.5-26B2024.12 | 79.1 | |
| InternVL2.5-78BModel Name=InternVL2.5-78B2024.12 | 78.8 | |
| InternVL2.5-8BModel Name=InternVL2.5-8B2024.12 | 78.4 | |
| InternVL2.5-38BModel Name=InternVL2.5-38B2024.12 | 78.3 | |
| Qwen3-VL + DISCO (10K)Backbone=Qwen3-VL, Alignment Strategy=DISCO (10K)2026.02 | 77.92 | |
| Qwen3-VL + Textual (10K)Backbone=Qwen3-VL, Alignment Strategy=Textual (10K)2026.02 | 77.85 | |
| Qwen3-VLBackbone=Qwen3-VL2026.02 | 77.68 | |
| InternVL2-40BModel Name=InternVL2-40B2024.12 | 77.6 | |
| InternVL2-Llama3-76BModel Name=InternVL2-Llama3-76B2024.12 | 77.6 | |
| GPT-4o-20240513Model Name=GPT-4o-202405132024.12 | 76.6 | |
| MiniCPM-v2.6 + HIPPOBackbone=MiniCPM-v2.6, Alignment Strategy=HIPPO2026.02 | 76.37 | |
| MiniCPM-v2.6Backbone=MiniCPM-v2.62026.02 | 76.32 | |
| InternVL2-8BModel Name=InternVL2-8B2024.12 | 75.8 | |
| InternVL2-26BModel Name=InternVL2-26B2024.12 | 75.6 | |
| InternVL2.5-4BModel Name=InternVL2.5-4B2024.12 | 75.5 | |
| InternVL-Chat-V1.5Model Name=InternVL-Chat-V1.52024.12 | 75.4 | |
| MiniCPM-V2.6Model Name=MiniCPM-V2.62024.12 | 75.2 | |
| Qwen2-VL-7BModel Name=Qwen2-VL-7B2024.12 | 74.4 | |
| InternVL2-4BModel Name=InternVL2-4B2024.12 | 71.1 | |
| InternVL2.5-2BModel Name=InternVL2.5-2B2024.12 | 70.2 | |
| InternVL2-2BModel Name=InternVL2-2B2024.12 | 66.3 | |
| InternVL2.5-1BModel Name=InternVL2.5-1B2024.12 | 60.9 | |
| InternVL2-1BModel Name=InternVL2-1B2024.12 | 57.5 |